In 1990, nine years after a study in Zambia to evaluate the efficacy of one, two, or three doses of hepatitis B vaccine, 101 of 255 recipients of a plasma-derived hepatitis B vaccine (Heptavax-B) were evaluated. The age at the time of the first vaccine dose ranged from <1 to 80 years (mean: 16; median: 10). Prior to the first vaccine dose, 70/101 had had no serological markers of exposure to the hepatitis B virus (HBV). In 1983, two years after the first vaccine dose, antibody to the hepatitis B surface antigen (anti-HBs) had been detectable in 90 of these 101 participants (89%). In 1990, anti-HBs was still detectable in 72/101 (71%), and was present at a protective level (greater than 10mIU) in 68/101 (67%). (In 1990, anti-HBs was still detectable in 52/70 (74%) who had had no HBV markers in 1981, and protective levels were detected in 47/70 [67%].) No difference was observed in protective levels of anti-HBs between recipients of two doses (23/31; 74%) or of three doses (26/36; 72%); however, the prevalence was lower among recipients of one dose (19/34; 56%). Between the years 1983-1990, hepatitis B virus (HBV) infections had occurred in 10/101 (10%); 8 of these 10 had received only one vaccine dose, and 6 had had detectable anti-HBs in 1983. (Three others had had acute HBV infections between 1981-1983 despite anti-HBs.)